USC Predict COVID Project
Predictive Epidemic Model for COVID-19 in Los Angeles County
University of Southern California, Department of Preventive Medicine
Epidemic model
Tables: Parameter estimates
\(R0\), \(r(t)\), \(\mu(t)\)
| mean (95% CI) | |
|---|---|
| R0 | 3.455 (3.336,3.656) |
| R(t) 2020-03-27 | 0.816 (0.733,0.904) |
| R(t) 2020-05-15 | 1.17 (1.032,1.268) |
| r(t) 2020-04-15 | 0.426 (0.173,0.685) |
| r(t) 2020-08-15 | 0.522 (0.184,0.735) |
\(\alpha(t)\), \(\kappa(t)\), \(\delta(t)\)
| mean (95% CI) Alpha_t | mean (95% CI) Kappa_t | mean (95% CI) Delta_t | |
|---|---|---|---|
| 2020-05-01 | 0.154 (0.127,0.175) | 0.57 (0.55,0.591) | 0.544 (0.525,0.565) |
| 2020-06-01 | 0.069 (0.043,0.087) | 0.59 (0.572,0.603) | 0.546 (0.52,0.569) |
CFR and IFR
| CFR mean (95% CI) | IFR mean (95% CI) | |
|---|---|---|
| 2020-03-01 | 0.0082 (0.0017,0.0163) | 0.0035 (0.0004,0.0079) |
| 2020-03-15 | 0.0082 (0.0051,0.0113) | 0.0036 (0.001,0.0067) |
| 2020-04-01 | 0.0149 (0.0123,0.0176) | 0.0064 (0.0022,0.011) |
| 2020-04-15 | 0.0257 (0.0221,0.0292) | 0.011 (0.0039,0.0187) |
| 2020-05-01 | 0.0334 (0.0291,0.0376) | 0.0143 (0.0051,0.0241) |
| 2020-05-15 | 0.0362 (0.0313,0.041) | 0.0155 (0.0055,0.0264) |
| 2020-06-01 | 0.0349 (0.0297,0.04) | 0.0149 (0.0054,0.0263) |
| 2020-06-15 | 0.0321 (0.0264,0.038) | 0.0137 (0.005,0.025) |
| 2020-07-01 | 0.0289 (0.0225,0.0359) | 0.0124 (0.0046,0.0236) |
| 2020-07-15 | 0.0268 (0.0197,0.0346) | 0.0115 (0.0041,0.0227) |
| 2020-08-01 | 0.0255 (0.0178,0.0338) | 0.0109 (0.0039,0.0221) |
| 2020-08-15 | 0.0253 (0.0172,0.0336) | 0.0108 (0.0039,0.0219) |
| 2020-09-01 | 0.0255 (0.0172,0.0336) | 0.0109 (0.004,0.0219) |
| 2020-09-15 | 0.0257 (0.0173,0.0337) | 0.011 (0.004,0.0219) |
| 2020-10-01 | 0.0259 (0.0174,0.0338) | 0.0111 (0.0041,0.0219) |
| 2020-10-15 | 0.0259 (0.0174,0.0337) | 0.0111 (0.0041,0.0219) |
| 2020-11-01 | 0.0256 (0.0169,0.0334) | 0.0109 (0.004,0.0218) |
| 2020-11-15 | 0.0252 (0.0165,0.0332) | 0.0108 (0.0039,0.0215) |
Figure: Model fits
Summarizes the epidemic model fit with COVID-19 data for LAC from March 1 through 2020-10-17 for all disease states across multiple views: New cases, representing new daily incidence; the current number in a compartment at a specific date, relevant for understanding current prevalence rates and comparing with healthcare capacity limitations; and cumulative counts until a specific date. Observed data for available compartments with cases attributable to nursing home residents removed, are plotted as black dots. The figure demonstrates that good model fits are achieved in all compartments across time.
- New = new daily incidence
- Current = current census in compartment
- Cumulative = running total over time
- Black dots depict COVID-19 data, with counts from outbreaks in nursing homes removed
- The dashed line represents healthcare resource capacity limits
Figures: Time-Varying Parameters
Reproductive Number \(R(t)\)
Fraction of Observed Infections \(r(t)\)
\(\alpha(t)\), \(\kappa(t)\), \(\delta(t)\)
Time-varying CFR(t)
Time-varying IFR(t)
Figures: Compartmental Variables
Current observed infections
Current estimated total infections
Current in hospital
New deaths
{-}
Risk model
Severe illness probabilities by profile
CFR(t) and IFR(t) by profile
Stage of disease by risk factor group
Estimated frequency of groups in the overall LAC population, and the population of individuals in each stage of disease from infected, hospitalized, admitted to ICU, to deceased on each date May.15, Aug.1, Oct.15 (2020): (A) Risk Groups, (B) age group, (C) obesity class groups, (D) any comorbidity group, and (E) smoking status group.
Risk group
Age
Cormorbidity
BMI
Smoking status
{-}